Controlling a DC inverter machine

[Xringer]

I guess people have been recording and playing back IR data bursts for some time.

ZipLabel.Com | Computerized Infrared Remote | Introduction

Ken Shirriff's blog: An Arduino universal remote: record and playback IR signals

It seems like a home automation system based on a PC that was on 24/7,
could be set up to replace a mini-split IR controller.

[S-F]

Originally Posted by Xringer

I guess people have been recording and playing back IR data bursts for some time.

ZipLabel.Com | Computerized Infrared Remote | Introduction

Ken Shirriff's blog: An Arduino universal remote: record and playback IR signals

It seems like a home automation system based on a PC that was on 24/7,
could be set up to replace a mini-split IR controller.

If you dig a little deeper you'll find that it's not so simple. I have a system that's on 24x7 and there are enormous obstacles in controlling a mini split. There is a form of handshake between the remote and the head unit.

[Xringer]

Originally Posted by S-F

If you dig a little deeper you'll find that it's not so simple. I have a system that's on 24x7 and there are enormous obstacles in controlling a mini split. There is a form of handshake between the remote and the head unit.

There is feedback. When the room temperature changes,
the remote sends out a data stream to tell the system to take a break, or work harder..

But, my understanding of the Sanyo remote, the data flow is one-way.
The remote sends out the stream and the system acts per it's firmware.

My Sanyo remote sends IR, whenever I push a command button,
(vent angle control, Ions, fan control, temp change & etc) or every 5 minutes.

In the event of a short grid failure, the Sanyo will come back on and shut down.
But, within 5 minutes, the IR remote will sent it's bust and return the System to it's mode, prior to the grid failure.

If the remote's battery dies, the IDU will switch to using it's built-in thermostat.
But, if there is a short grid fail, the system is going to stay turned off..
Which means that the Fall season is a good time to replace the batteries..



Anyways, I was thinking if you could record all your standard command busts on a PC disk,
then your control program could just pick out the right file and transmit it with an IR LED.
Of course, the PC would need a room temperature sensor.

And the remote control would have to be blocked from interfering..

I'm pretty sure that I could get a wired thermostat for my model Sanyo.
It would require installing a wiring harness inside the IDU..
But, since I have a remote, (and a back-up) I don't see the need for a wired stat..

One other way to intercede into link using the remote,
is to add a temperature change generator to the remote's sensor.
A simple device to trick the remote into thinking the room temperature
had changed, would cause it to send new commands to the IDU..

Like if you had a solar GTI and you wanted to AC to really crank hard after 10AM on a sunny day,
you could place the remote on the rug, in a spot that starts getting sunshine at 10AM..

[Acuario]

Originally Posted by AC_Hacker

Do you have a schematic for your interface board?

I'd like to see if I can make some sense out of my Sanyo.

Thanks,

-AC

Here it is. The device that interfaces from the TTL logic to the power leads is the HF6 3EB - probably a custom device as I can't find any reference to it anywhere. The rest is just basic RS232 to TTL level conversion stuff.

The CUT is where I break the comms from the internal unit to be able to send my own commands.

The CLK signal from the HF6 doesn't appear to be doing much as the data isn't synchronised to it in any way. It runs at 100Hz. I may be wrong but for the moment I'm not worrying about it.

Nigel

[S-F]

Xringer, if you aren't going to dig deeper like I said you're just going to have to trust me when I say that it's almost impossible to control a mini split with an IR blaster. Both AC and IR blaster manufacturers have been talking about this for years and to my knowledge no one has managed to get it to work.

[Xringer]

It's a short burst.. Maybe running at 44khz, giving it a lot of data..
Here's an HD video.. (17.5 mb)
Sanyo ASHP IR data video by Xringer - Photobucket

Maybe I can look at it in slow mo and see more?

[Acuario]

Originally Posted by S-F

Xringer, if you aren't going to dig deeper like I said you're just going to have to trust me when I say that it's almost impossible to control a mini split with an IR blaster. Both AC and IR blaster manufacturers have been talking about this for years and to my knowledge no one has managed to get it to work.

Although it's not the method I'm using I would tend to agree with S-F. In the thread you pointed us to on refrigeration-engineer the reason he didn't follow that route was lack of feedback. The problem with remote controls is they are one-way only so yes, I'm sure you can emulate some functions but you can never guarantee the status of the machine before sending the command. The only thing you could assume is that it is on so send the command to turn it off and start from zero.

The other problem with IR is it's light so if the remote isn't in view of the machine it won't operate it. There are machines that have a mode where the thermometer is in the handset and they tell the machine what the temperature is where the remote is placed, thereby controlling the machine, but I guess they must have some fallback in case of communication failure (as you described Xringer).

[Acuario]

Originally Posted by Xringer

It's a short burst.. Maybe running at 44khz, giving it a lot of data..
Here's an HD video.. (17.5 mb)
Sanyo ASHP IR data video by Xringer - Photobucket

Maybe I can look at it in slow mo and see more?

You'ld be very lucky to see what is going on - an oscilloscope or logic analyser are the only way. You'll then see something akin to the picture below. From there you can start analysing the pulse sequence/s.

[AC_Hacker]

So I thought I'd try to find out what I can about my 9K BTU sanyo mini-split...


This is the indoor unit with the air-filter shroud removed...

This is the diagram on the wiring cover...

Here's a close up of the wires...

This is the DC voltage measured from ground to the data wire...

This is the AC voltage measured from ground to the data wire...

Question for Nigel... what frequency is required to be able to analyze the data pckets?

-AC

[Acuario]

The only answer I can give you is that it could be absolutely anything.

What you need to do is open the machine up and find out where the data wire goes/what it connects to and based on that you should be able to drawn out a bit of the circuit at one or other of the ends. Then you need to know what voltage it's all working at; normally it will eventually reach TTL levels (0V low to +5V high). Then connect an oscilloscope or logic analyzer, capture the data trace and from there you can deduce the operating frequency and from there the baud rate of the data.

If you're lucky it will (as in my case) turn out to be a standard RS232 speed, you may need to analyse the data frame to determine what the data structure is, stop bits, parity bits etc. as there are many possible combinations. In my case the Daikin was 1 start bit, 2 stop bits and odd parity and running at 1200 baud.

A cheap logic analyzer is probably the easiest way to capture the data frames for analysis - you can find them on ebay.

It's a challenge but can be done. If you have no electronics/software (you will need this next) knowledge then you'll probably need to enlist the help of someone who does.

Nigel